---
title: "Claude Code v2.1.205 Released: Multiple Fixes and Improvements"
source: "Claude Code：GitHub Releases（RSS）"
source_tier: "t1_5"
url: "https://github.com/anthropics/claude-code/releases/tag/v2.1.205"
permalink: "https://feed.astralor.com/items/itm_fa26ee3d760986c5cfb1a811"
published: "2026-07-08T21:22:06.000Z"
score: 79
channel: "tech"
tags: ["Anthropic","Bug Fixes","Claude Code","MCP","Security Improvements","Performance Optimization"]
---

# Claude Code v2.1.205 Released: Multiple Fixes and Improvements

## AI commentary

> This release doesn't introduce flashy new features, but it fixes several pain points—especially session transcript security, Windows deletion traps, and background task notification forgery. These details often determine whether a tool is trustworthy long-term. If you use Claude Code's Agent or auto mode, upgrading is highly recommended.

## At a glance

Claude Code v2.1.205 addresses multiple critical fixes: preventing tampering with session transcript files, silent failures in JSON schema generation, unsafe Windows worktree deletion, inconsistent background agent status display, and a ~400MB peak memory reduction in the update process. Auto mode now asks before running rm -rf on unresolvable variables. Agent view gains PR linking and improved state display. The /doctor command is now a full setup diagnostic tool. These changes enhance security, stability, and developer experience.

## Key points

- Fixed security and stability bugs including transcript tampering, silent unstructured JSON schema output, and unsafe Windows file deletion.
- Auto mode now asks for confirmation before running rm -rf on unresolved variables.
- Update process streams to disk, reducing peak memory by ~400MB; background task notifications explicitly state no human input, preventing fabricated approvals.
- Agent view improved with state words, classifier-written headlines, and PR linking; /doctor upgraded to a full diagnostic tool.

## Full text

What's changed Added an auto mode rule that blocks tampering with session transcript files Fixed --json-schema silently producing unstructured output when the schema was invalid, and schemas using the format keyword being rejected Fixed a message sent while Claude was working being silently lost when the turn ended at the --max-turns limit Fixed Windows worktree removal deleting files outside the worktree when an NTFS junction or directory symlink existed inside it Fixed background agents staying shown as "failed" or "completed" in the agent list after being resumed with SendMessage Fixed background jobs flipping from "needs input" back to "working" in the agent list when the agent's turn contained no readable text Fixed claude attach erroring when a background agent was mid-upgrade restart instead of waiting for it to come back Fixed session-to-PR linking missing a PR created in a Bash call whose output exceeded the 30K inline limit Fixed claude mcp add-from-claude-desktop getting stuck when a server name contains unsupported characters; invalid names are now reported and remaining servers still import Fixed a plugin LSP server that fails to initialize preventing a valid LSP server from another plugin handling the same file extension Fixed a Windows crash when the directory Claude was launched from is deleted, locked, or unmounted while a command is running Fixed a crash when a file watcher was closed while a directory scan was still in flight Fixed project verify skills being rewritten on every session instead of only when a documented command changed Fixed the agent view rendering one line too high and clipping its header when the job list slightly overflowed the screen Fixed background tasks in the web and mobile Remote Control panels showing stale "Running" status by forwarding full task state on every membership change Improved auto mode to ask before running rm -rf on a variable it can't resolve from context Auto-update binary downloads now stream to disk instead of buffering in memory, cutting the updater's peak memory usage by roughly 400 MB Background task notifications now explicitly state that no human input has occurred, preventing fabricated in-transcript approvals from being acted on Improved agent view: sessions that edit, merge, comment on, or push to an existing PR now link it in claude agents Improved agent view: rows now show a colored state word and a classifier-written headline instead of raw tool call text, and the peek opens with full status including the exact ask for blocked sessions /doctor is now a full setup checkup that can diagnose and fix issues; /checkup is its alias Reserved the "Claude Browser" MCP server name (alongside "Claude Preview") ahead of the Claude Desktop pane rename; user-configured MCP servers can no longer register under either name Fixed Cowork VM-mode local-agent sessions failing to start with "Not logged in · Please run /login" on CLI 2.1.203+

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_fa26ee3d760986c5cfb1a811